- Email: [email protected]
Outcome of repeated radiosurgery for recurrent metastatic brain tumors
Clinical Neurology and Neurosurgery 109 (2007) 132–137
Outcome of repeated radiosurgery for recurrent metastatic brain tumors Ki-Young Kwon, Doo-Sik Kong, Jung-Il Lee ∗ , Do-Hyun Nam, Kwan Park, Jong Hyun Kim Department of Neurosurgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-gu, Seoul 135-710, Republic of Korea Received 5 February 2006; received in revised form 19 June 2006; accepted 24 June 2006
Abstract Objective: We investigated the outcome of repeated gamma knife radiosurgery (GKS) for local or remote recurrence after initial radiosurgery. Material and methods: We retrospectively reviewed 204 patients who were treated with GKS. Among them 43 patients (21%) underwent GKS more than once. The second GKS was given for recurrence at the previously treated sites in 16 patients, new lesions at remote sites in 13, and both local recurrence and new lesions in 14. Results: The median survival from the first GKS was 36 (7–190) weeks in all patients and 68 (16–156) weeks in 43 patients with repeated GKS. The median time from the first GKS to the second was 37 weeks. The median survival from the second radiosurgical intervention was 32 (7–132) weeks. Local control rate at 6 months after salvage GKS was 90.7%. RPA class was the commonly dominant prognostic factor in both initial and salvage GKS. Conclusion: Recurrence is common for patients with metastatic brain tumors after initial radiosurgery. Local control and survival time after salvage treatment are comparable with those after initial radiosurgery. GKS as a salvage treatment may provide additional survival benefit in selected patients. © 2006 Elsevier B.V. All rights reserved. Keywords: Metastatic brain tumor; Gamma knife; Radiosurgery; Salvage therapy
1. Introduction Radiosurgery can be used as a primary treatment for metastatic disease of the brain [1–5]. Using this technique, multiple metastases can be treated with a single procedure and lesions with surgically inaccessible location can be treated without difficulty [6–9]. Radiosurgery has the advantages of high safety and low cost in comparison with open surgical treatment. Regarding overall survival, radiosurgery for metastatic disease of the brain is comparable with open surgical treatment. However, it is not rare for patients to present with local recurrence or new lesions after either open surgical treatment or radiosurgery. Another advantage of radiosurgery is that repeated procedure for recurrent or progressive lesions ∗
Corresponding author. Tel.: +82 2 3410 3499/94; fax: +82 2 3410 0048. E-mail address: [email protected] (J.-I. Lee).
0303-8467/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.clineuro.2006.06.007
is relatively easier and safer than other treatment modalities. The benefit of salvage therapy for progressive lesions after initial radiosurgical treatment has not been defined well. In this study, we investigated the outcome of salvage gamma knife radiosurgery (GKS) for local or remote recurrence after initial radiosurgery.
2. Materials and methods 2.1. Patient population Between January 2002 and October 2005, 204 patients (110 male and 94 female) with 847 metastatic brain lesions were treated with GKS at our institution. The mean age of the patients was 55.9 (23–87) years. The common sites of the primary tumors in the order of frequency were the lung in 93
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
(46%) patients, breast in 26 (13%), kidney in 20 (10%), colorectum in 14 (7%) and others (25%). The treatment given in addition to GKS included the surgical removal of tumors in 32 patients (16%) and whole brain radiotherapy (WBRT) in 135 (67%). The Karnofsky performance status (KPS) score at the time of initial GKS was 70 or more in 139 (68%) patients, and less than 70 in 65 (32%). Extracranial metastases were present in 106 (52%) patients and absent in 98 (48%). Primary site tumor was progressive in 91 (45%) and stable or controlled in 113(55%). According to the recursive partitioning analysis (RPA) classification proposed by the radiation therapy oncology group (RTOG) [10], 44 (22%) patients belonged to RPA class I (KPS score of 70 or more, less than 65 years old, no extracranial metastasis, and controlled primary disease), 95 (47%) patients belonged to RPA class II (patients not in classes I or III), and 65 (32%) patients belonged to RPA class III (KPS score of less than 70) (Table 1). Follow-up imaging studies were recommended at every 3 months after treatment or at the time of developing neurologic symptoms. Progression of a treated lesion was defined as an increase in longest diameter more than 20% on sequential images [11]. Among the 204 patients, 43 (21%) underwent GKS more than once. The number of GKS procedures was 2 in 32 patients (75%), and more than 2 in 11 (25%). At the time of the second Table 1 Clinical characteristics of patients in entire group
133
radiosurgical treatment, 16 patients (36%) had recurrence at the previously treated sites, 13 (30%) at remote sites, and 14 (34%) at both the previously treated and remote sites. Distribution of the patients according to RPA class was 7 (16%), 25 (58%), and 11(26%) for classes I–III. 2.2. Radiosurgical technique Leksell Gamma Knife types B and C (Elekta Instruments, Inc., Atlanta, GA) were used to administer the treatment. The stereotactic frame was applied to the patient under local anesthesia. Tumors were localized using high-resolution contrastenhanced magnetic resonance imaging. Treatment planning was performed using Gamma Plan software Version 5.32 and 5.34 (Elekta Instruments, Inc., Atlanta, GA). The mean total tumor volume at initial GKS was 9.0 cm3 (0.1–104), and the mean number of lesions per patient was 4.2 (1–30). Mean volume of individual lesion was 1.9 cm3 (0.1–48.7). The mean marginal dose was 16.1 Gy (10–30) with a median marginal isodose of 50% (35–65%). Dose selection was based on various factors including the volume and number of the lesion, location, prior radiation therapy, and a predicted dose–response relationship for brain parenchymal necrosis. The most important factor considered in dose prescription was individual or total tumor volume. For the patients with single lesion, dose prescription was performed according to RTOG report [12] with a little modification considering other factors. In the patients with multiple lesions, marginal dose was modified with consideration of total tumor volume as well as volume of the individual lesion. In repeated radiosurgery marginal dose was reduced about 25% from the dose used for the same volume in initial treatment. The mean total tumor volume and mean number of lesions at the second GKS were 7.6 cm3 (0.8–25) and 4.5 cm3 (1–22), respectively. The mean volume of the individual lesion was 2.2 cm3 (0.1–20.5) and the mean marginal dose was 15.3 Gy (12–22).
Variables
Number of patients
Age
Mean age 55.9 (range 23–78)
Sex M F
110 (54%) 94 (46%)
Primary site Lung Breast Kidney Colorectum Others
93 (46%) 26 (13%) 20 (10%) 14 (7%) 51 (25%)
WBRT Yes No
135 (66%) 69 (34%)
RPA class I II III
44 (22%) 95 (47%) 65 (32%)
Number of lesion 1–3 >3
134 (66%) 70 (34%)
Statistical analysis was performed using commercial software SPSS Version 11.0 (SPSS Inc., Chicago, IL). The Kaplan–Meier method was used to calculate the patient survival period and actuarial survival rate. The prognostic factors were analyzed for significance using univariate analysis with log-rank tests. Subsequently, a multivariate analysis was performed using Cox proportional hazard regression analysis.
Tumor volume (cm3 ) <8 >8
122 (60%) 82 (40%)
3. Results
Extracranial metastasis Yes No Primary tumor controlled Yes No
2.3. Statistical analysis
106 (52%) 98 (48%) 113 (55%) 91 (45%)
3.1. Clinical outcome in the entire group The mean follow up period was 38 weeks (4–182). A hundred fifty four patients died and 50 patients were still alive at the time of the last follow-up. The median survival of all
134
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
Fig. 1. Kaplan–Meier survival curve showing overall survival for the entire group after first radiosurgery.
the patients after initial radiosurgery was 36 (4–132) weeks (Fig. 1). The actuarial survival rate was 54% at 6 months and 32% at 1 year. The local control rate and rate without development of new lesions after the first GKS were 94.1% and 89.2% at 6 months and 86.2% and 78.3% at 1 year, respectively. The cause of death was determined in 123 patients, but was indeterminate in 31 patients. Eighty-five patients (55%) died of systemic causes, 38 (25%) of neurologic causes, and 31(25%) of unknown causes. Symptomatic radiation necrosis developed in 3 (1.5%) of the 204 patients after the first GKS. They had undergone prior WBRT with total dose of 30 Gy in 10 fractions. Tumor volume and marginal dose (cm3 /marginal dose, Gy) at the sites of radiation necrosis was 7.3/15, 4.5/20, and 3.0/25, each respectively. 3.2. Prognostic factors in the entire group Primary site, number of metastatic lesions, total tumor volume, WBRT and RPA class were included as possible prognostic factors for survival in the univariate analysis. The results are summarized in Table 2. Primary site was not a significant factor influencing survival. Number of lesions 3 or less (p = 0.001), total tumor volume 8 cm3 or less (p = 0.008), better RPA class (P < 0.001), and combined WBRT (p = 0.005) were related to longer survival. Multivariate analysis revealed that RPA class was the most dominant prognostic factors affecting survival (median survival of 64, 43, and 16 weeks in RPA classes I–III, respectively, p < 0.001) (Fig. 2).
Fig. 2. Kaplan–Meier survival curve showing overall survival stratified by recursive partitioning analysis class (A) and number of lesions (B) for all patients after radiosurgery.
3.3. Clinical outcome in salvage GKS group A subset of 43 patients (21% of 204 patients) underwent repeated GKS for local or distant recurrence (Table 3). The probability of additional procedures increased with longer survival and was 35% for those who survive longer than 6 months after the initial treatment. The mean follow-up period after the second GKS was 38 weeks (7–156). Thirty-three of these patients died and 10 were still alive at the time of the last follow-up. The mean interval between the first GKS and salvage GKS was 37 (7–80) weeks. The median survival from the first GKS was 68 (16–156) weeks in this patients group and that after the second GKS was 32 (7–132) weeks (Fig. 3). The cause of death was determined in 30 patients, but was indeterminate in three patients. Sixteen (48.4%) patients died of systemic causes and 14 (42.4%) died of neurologic causes.
Table 2 Uni- and multivariate analysis of prognostic factors for all patients Variables
Median survival (weeks)
Univariate (p-value)
Multivariate (p-value)
Primary site (lung vs. non-lung) No. of lesion (≤3 vs. >3) Tumor volume (≤8 cm3 vs. >8 cm3 ) Presence of WBRT (yes vs. no) RPA classes (I, II vs. III)
40 vs. 28 49 vs. 20 45 vs. 22 45 vs. 23 64, 43 vs. 16
0.378 0.001 0.008 0.005 <0.001
EA 0.009 0.068 0.029 <0.001
WBRT: whole brain radiotherapy; RPA: recursive partitioning analysis; EA: excluded from analysis.
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
135
Table 3 Clinical characteristics of patients in salvage radiosurgery group Variables
Number of patients
Age
Mean age 54.7 (range 35–75)
Sex M F
25 (58%) 18 (42%)
Primary site Lung Breast Kidney Others
22 (51%) 5 (12%) 7 (16%) 9 (21%)
RPA class I II III
7 (16%) 25 (58%) 11 (26%)
Number of lesion 1–3 >3
21 (49%) 22 (51%)
Tumor volume (cm3 ) <8 >8
21 (49%) 22 (51%)
Type of recurrence Local recurrence New lesion at remote site Both
16 (37%) 13 (30%) 14 (33%)
Fig. 4. Kaplan–Meier survival curve showing overall survival stratified by recursive partitioning analysis class for salvage radiosurgery group.
tumor volume change in the patients with repeated radiosurgery showed that radiosurgery for local recurrence was effective as well as initial radiosurgery in newly diagnosed patients (Fig. 5). Symptomatic radiation necrosis developed in eight patients (18.6%) after the second GKS. Mean tumor volume/marginal dose at the site of radiation necrosis was 5.8 cm3 (1.2–12.2)/19 Gy (17–20 Gy) at the first radiosurgery and 2.3 cm3 (0.2–8.4), 17 Gy (13–20) at the second radiosurgery. 3.4. Prognostic factors in salvage GKS group In the univariate analyses, interval between salvage GKS and the initial GKS, number of lesions, tumor volume, and type of recurrence (local progression or development of new lesions) were not significant factors influencing survival (Table 4). There was a significant difference in the median survival in relation to RPA class (median survival of RPA I–III was 42, 36 and 12 weeks, p = 0.007) (Fig. 4). In the multivariate analyses, the RPA class was the only statistically significant factor (p = 0.015).
Fig. 3. Kaplan–Meier survival curve showing overall survival for salvage radiosurgery group after second radiosurgery.
The actuarial survival rate after the second GKS was 57.6% at 6 months and 28% at 1 year. The local control rate and rate without development of new lesions at 6 months after the second GKS were 90.7% and 86%, respectively. The
4. Discussion Radiosurgery is recognized as a powerful treatment modality for selected patients with brain metastases [13]. It was reported that the overall median survivals ranged from 24 to 44 weeks in patients treated with radiosurgery for brain metastases of all origins [4,9,18–21]. Also, it was reported
Table 4 Uni- and multivariate analysis of prognostic factors for salvage radiosurgery group Variables
Median survival (weeks)
Univariate (p-value)
Multivariate (p-value)
Number of lesion (≤3 vs. >3) Tumor volume (≤8 cm3 vs. >8 cm3 ) Interval between the first and second GKS (≥35 weeks vs. <35 weeks) Type of recurrence (local, new lesion vs. both) RPA classes (I, II vs. III)
36 vs. 28 36 vs. 28 36 vs. 28 36, 26 vs. 32 42, 36 vs. 12
0.725 0.074 0.650 0.858 0.007
0.526 0.657 0.424 EA 0.015
RPA: recursive partitioning analysis; EA: excluded from analysis.
136
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
to be beneficial in the end-stage management of carefully selected patients with numerous brain metastases, even in those with 10 or more lesions [13–17]. However, little is known about the outcome of radiosurgery as a salvage therapy. Chen et al. [6] reported that radiosurgical salvage therapy extended the survival time. They analyzed the patients who underwent salvage radiosurgery for newly developed lesions after initial radiosurgery. In their series, the median time from the first radiosurgical procedure to the first salvage treatment was 17.4 weeks and the median survival from the second radiosurgical intervention was 28 weeks. Our study shows somewhat longer interval of 37 weeks from the first radiosurgical procedure to the first salvage treatment and a comparable median survival of 32 weeks from the second radiosurgical intervention. The median survival after the salvage GKS (32 weeks) is comparable with the overall median survival from the first GKS in all patients (36 weeks). Also the median survival after the salvage GKS for each RPA class is similar to the reported outcome of radiosurgery alone for newly diagnosed brain metastasis in the literature [22], in which the median survival for patients of RPA classes I–III were 13.4, 9.3, and 1.5 months. The median survival of the salvage treatment group from the first procedure (68 weeks) is significantly longer than that of entire group (36 weeks). One possible explanation is that the patients in the salvage treatment group have favorable prognostic factors and live long enough to experience recurrence and become candidates for repeat radiosurgery. Such a self-selection process may be an important cause of the observed longer overall survival time. However, our data also provide a reasonable interpretation for the beneficial effect of salvage GKS. Because the probability of local tumor control after salvage GKS is comparable to that after initial GKS, it is thought that multiple radiosurgery procedures result in actual prolongation of freedom from progression of both already treated lesions (shown in Fig. 5) and new lesions, thereby extending the lifetime of the patient.
Though direct comparison with other treatment is difficult due to limited data, the survival after the salvage radiosurgery in our study appears to be better than reported outcome after other method of reirradiation. In a report of whole or partial brain reirradiation [23], in which median dose of 20 Gy in 10 fractions was given for progressive brain metastases after the initial course of WBRT, overall median survival was 4 months. Though it seemed to include more patients with poor performance status at the time of reirradiation (ECOG performance status of 3 or worse in 49%) than ours at the time of the second GKS (RPA class 3 in 24%), better ECOG performance status was not associated with improved survival in that study. Overall median survival of 4 months is comparable to that of RPA class III (12 weeks) in our study and much shorter than those of RPA I (42 weeks) and II (36 weeks). In our study, the common significant prognostic factor affecting survival after the initial and salvage GKS is RPA class. It is thought that performance status becomes more influential on survival when control of brain lesion is effective. With respect to the influence of the number of lesions on survival, other authors could find no effect when treating up to four to five tumors [24–26]. The number of lesions shows a wider variation in our series, and the number of lesions or lesion volume affected survival. However, RPA class gave the most dominant influence. Therefore, it is suggested that a limited number of prognostic factors, such as the KPS score or RPA class at the time of the procedure, constitute the major determinants of outcome in both salvage GKS and initial GKS, and that the decision for additional radiosurgery should be made on the consideration of these predominant factors. Also, it should be noted that situation requiring multiple GKS procedures is not exceptional, but rather common and predictable in the course of treating metastatic brain lesions. The incidence of complications tends to increase after multiple GKS procedures, however, it seems that this risk is outweighed by the expectable prolongation of survival. Additional benefit of repeated radiosurgery may be an improvement in the quality of life, however, it is beyond the collected data of this study and further prospective trials are needed to clarify this issue.
5. Conclusion
Fig. 5. Graph depicting change of tumor volume in patients with locally recurrent lesions after the first and second GKS. Data are from 16 patients in whom MR images were obtained at 3 months after the second GKS. The volume of recurrent lesions was stable or decreased in 13 patients and increased in three after repeated treatment. Tumor volume is presented with mean percentage of the volume at the first GKS and standard deviation. * Percentage volume at the time of maximal reduction during serial followup.
Local or remote recurrence which needs salvage treatment is common for patients with metastatic brain tumors after initial radiosurgery. Repeated GKS is effective in treating locally progressive or new lesions after initial GKS. If the major prognostic factors such as RPA class are favorable, repeated radiosurgical procedure may be beneficial for an increase in survival as like initial radiosurgery.
Acknowledgment This work was supported by a grant of Insung Foundation for Medical Research (J.-I. Lee).
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
References [1] Adler JR, Cox RS, Kaplan I, Martin DP. Stereotactic radiosurgical treatment of brain metastases. J Neurosurg 1992;76:444–9. [2] Alexander III E, Moriarty TM, Davis RB, Wen PY, Fine HA, Black PM, et al. Stereotactic radiosurgery for the definitive, noninvasive treatment of brain metastases. J Natl Cancer Inst 1995;87:34–40. [3] Auchter RM, Lamond JP, Alexander E, Buatti JM, Chappell R, Friedman WA, et al. A multiinstitutional outcome and prognostic factor analysis of radiosurgery for resectable single brain metastasis. Int J Radiat Oncol Biol Phys 1996;35:27–35. [4] Flickinger JC, Kondziolka D, Lunsford LD, Coffey RJ, Goodman ML, Shaw EG, et al. A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys 1994;28:797–802. [5] Shirato H, Takamura A, Tomita M, Suzuki K, Nishioka T, Isu T, et al. Stereotactic irradiation without whole-brain irradiation for single brain metastasis. Int J Radiat Oncol Biol Phys 1997;37:385–91. [6] Chen JCT, Petrovich Z, Giannotta SL, Yu C, Apuzzo ML. Radiosurgery salvage therapy for patients presenting with recurrence of metastatic disease to the brain. Neurosurgery 2000;46:860–7. [7] Chen JCT, Petrovich Z, O’Day S, Morton D, Essner R, Giannotta SL, et al. Stereotactic radiosurgery in the treatment of metastatic disease to brain. Neurosurgery 2000;47:268–81. [8] Gerosa M, Nicolato A, Foroni R, Zanotti B, Tomazzoli L, Miscusi M, et al. Gamma knife radiosurgery for brain metastases: a primary therapeutic option. J Neurosurg 2000;97(Suppl 5):515–24. [9] Kim YS, Kondziolka D, Flickinger JC, Lunsford LD. Stereotactic radiosurgery for patients with nonsmall cell lung carcinoma metastatic to the brain. Cancer 1997;80:2075–83. [10] Gaspar L, Scott C, Rotman M, Asbell S, Phillips T. Recursive partitioning analysis of prognostic factors in three radiation therapy oncology group brain metastases trials. Int J Radiat Oncol Biol Phys 1997;37(4):745–51. [11] Therasse P, Arbuck SG, Eisenhauer EA, Wanders J. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 2000;92(3):205–16. [12] Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90-05. Int J Radiat Oncol Biol Phys 2000;47(2):291–8. [13] Yamamoto M, Ide M, Nishio. Gamma knife radiosurgery for numerous brain metastases: is this a safe treatment? Int J Radiat Oncol Biol Phys 2002;53(5):1279–83.
137
[14] Suzuki S, Omagari J, Nishio S, Nishiye E, Fukui M. Gamma knife radiosurgery for simultaneous multiple metastatic tumors. J Neurosurg 2000;93(Suppl 3):30–1. [15] Nam TK, Lee JI, Jung YJ, Im YS, An HY, Nam DH, et al. Gamma knife surgery for brain metastases in patients harboring four or more lesions: survival and prognostic factors. J Neurosurg 2005;(Suppl 102):147– 50. [16] Yamamoto H, Takagi K, Ushio T. Gamma knife radiosurgery for diffuse brain metastases. Noushinkeigeka Sokuho 1998;8:391–8. [17] Yamamoto M, Ide M, Jimbo M. Gamma knife radiosurgery with numerous target points for intracranially disseminated metastases: early experience in 3 patients and experimental analysis of whole brain irradiation doses. Radiosurgery 1997:94–109. [18] Engenhart R, Kimmig BN, Hover KH, Hover KH, Wowra B, Romahn J, et al. Long-term follow-up for brain metastases treated by percutaneous stereotactic single high-dose irradiation. Cancer 1993;71:1353– 61. [19] Kihlstrom L, Karlsson B, Lindquist C. Gamma knife surgery for cerebral metastases. Implications for survival based on 16 year experience. Stereotact Funct Neurosurg 1993;61(Suppl 1):45–50. [20] Kim DG, Chung HT, Gwak HS, Paek SH, Jung HW, Han DH. Gamma knife radiosurgery for brain metastases: prognostic factors for survival and local control. J Neurosurg 2000;93(Suppl 3):23– 9. [21] Somaza S, Kondziolka D, Lunsford LD, Kirkwood JM, Flickinger JC. Stereotactic radiosurgery for cerebral metastatic melanoma. J Neurosurg 1993;79:661–6. [22] Lutterbach J, Cyron D, kenne K, Ostertag CB. Radiosurgery followed by planned observation in patients with one to three brain metastases. Neurosurgery 2003;52:1066–74. [23] Wong WW, Schild SE, Sawyer TE. Analysis of outcome in patients reirradiated for brain metastases. Int J Radiat Oncol Biol Phys 1996;34(3):585–90. [24] Chen J, O’Day S, Morton D, Essner R, Cohen-Gadol A, Macpherson D, et al. Stereotactic radiosurgery in the treatment of metastatic disease to the brain. Neurosurgery 2000;47(2):268–79. [25] Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC. Stereotactic radiosurgery plus whole radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat Oncol Biol Phys 1999;45:427–34. [26] Sneed P, Lamborn K, Forstner J, McDermott M, Chang S, Park E, et al. Radiosurgery for brain metastases: Is whole brain radiotherapy necessary? Int J Radiat Oncol Biol Phys 1999;43:549–58.
Outcome of repeated radiosurgery for recurrent metastatic brain tumors Ki-Young Kwon, Doo-Sik Kong, Jung-Il Lee ∗ , Do-Hyun Nam, Kwan Park, Jong Hyun Kim Department of Neurosurgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Kangnam-gu, Seoul 135-710, Republic of Korea Received 5 February 2006; received in revised form 19 June 2006; accepted 24 June 2006
Abstract Objective: We investigated the outcome of repeated gamma knife radiosurgery (GKS) for local or remote recurrence after initial radiosurgery. Material and methods: We retrospectively reviewed 204 patients who were treated with GKS. Among them 43 patients (21%) underwent GKS more than once. The second GKS was given for recurrence at the previously treated sites in 16 patients, new lesions at remote sites in 13, and both local recurrence and new lesions in 14. Results: The median survival from the first GKS was 36 (7–190) weeks in all patients and 68 (16–156) weeks in 43 patients with repeated GKS. The median time from the first GKS to the second was 37 weeks. The median survival from the second radiosurgical intervention was 32 (7–132) weeks. Local control rate at 6 months after salvage GKS was 90.7%. RPA class was the commonly dominant prognostic factor in both initial and salvage GKS. Conclusion: Recurrence is common for patients with metastatic brain tumors after initial radiosurgery. Local control and survival time after salvage treatment are comparable with those after initial radiosurgery. GKS as a salvage treatment may provide additional survival benefit in selected patients. © 2006 Elsevier B.V. All rights reserved. Keywords: Metastatic brain tumor; Gamma knife; Radiosurgery; Salvage therapy
1. Introduction Radiosurgery can be used as a primary treatment for metastatic disease of the brain [1–5]. Using this technique, multiple metastases can be treated with a single procedure and lesions with surgically inaccessible location can be treated without difficulty [6–9]. Radiosurgery has the advantages of high safety and low cost in comparison with open surgical treatment. Regarding overall survival, radiosurgery for metastatic disease of the brain is comparable with open surgical treatment. However, it is not rare for patients to present with local recurrence or new lesions after either open surgical treatment or radiosurgery. Another advantage of radiosurgery is that repeated procedure for recurrent or progressive lesions ∗
Corresponding author. Tel.: +82 2 3410 3499/94; fax: +82 2 3410 0048. E-mail address: [email protected] (J.-I. Lee).
0303-8467/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.clineuro.2006.06.007
is relatively easier and safer than other treatment modalities. The benefit of salvage therapy for progressive lesions after initial radiosurgical treatment has not been defined well. In this study, we investigated the outcome of salvage gamma knife radiosurgery (GKS) for local or remote recurrence after initial radiosurgery.
2. Materials and methods 2.1. Patient population Between January 2002 and October 2005, 204 patients (110 male and 94 female) with 847 metastatic brain lesions were treated with GKS at our institution. The mean age of the patients was 55.9 (23–87) years. The common sites of the primary tumors in the order of frequency were the lung in 93
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
(46%) patients, breast in 26 (13%), kidney in 20 (10%), colorectum in 14 (7%) and others (25%). The treatment given in addition to GKS included the surgical removal of tumors in 32 patients (16%) and whole brain radiotherapy (WBRT) in 135 (67%). The Karnofsky performance status (KPS) score at the time of initial GKS was 70 or more in 139 (68%) patients, and less than 70 in 65 (32%). Extracranial metastases were present in 106 (52%) patients and absent in 98 (48%). Primary site tumor was progressive in 91 (45%) and stable or controlled in 113(55%). According to the recursive partitioning analysis (RPA) classification proposed by the radiation therapy oncology group (RTOG) [10], 44 (22%) patients belonged to RPA class I (KPS score of 70 or more, less than 65 years old, no extracranial metastasis, and controlled primary disease), 95 (47%) patients belonged to RPA class II (patients not in classes I or III), and 65 (32%) patients belonged to RPA class III (KPS score of less than 70) (Table 1). Follow-up imaging studies were recommended at every 3 months after treatment or at the time of developing neurologic symptoms. Progression of a treated lesion was defined as an increase in longest diameter more than 20% on sequential images [11]. Among the 204 patients, 43 (21%) underwent GKS more than once. The number of GKS procedures was 2 in 32 patients (75%), and more than 2 in 11 (25%). At the time of the second Table 1 Clinical characteristics of patients in entire group
133
radiosurgical treatment, 16 patients (36%) had recurrence at the previously treated sites, 13 (30%) at remote sites, and 14 (34%) at both the previously treated and remote sites. Distribution of the patients according to RPA class was 7 (16%), 25 (58%), and 11(26%) for classes I–III. 2.2. Radiosurgical technique Leksell Gamma Knife types B and C (Elekta Instruments, Inc., Atlanta, GA) were used to administer the treatment. The stereotactic frame was applied to the patient under local anesthesia. Tumors were localized using high-resolution contrastenhanced magnetic resonance imaging. Treatment planning was performed using Gamma Plan software Version 5.32 and 5.34 (Elekta Instruments, Inc., Atlanta, GA). The mean total tumor volume at initial GKS was 9.0 cm3 (0.1–104), and the mean number of lesions per patient was 4.2 (1–30). Mean volume of individual lesion was 1.9 cm3 (0.1–48.7). The mean marginal dose was 16.1 Gy (10–30) with a median marginal isodose of 50% (35–65%). Dose selection was based on various factors including the volume and number of the lesion, location, prior radiation therapy, and a predicted dose–response relationship for brain parenchymal necrosis. The most important factor considered in dose prescription was individual or total tumor volume. For the patients with single lesion, dose prescription was performed according to RTOG report [12] with a little modification considering other factors. In the patients with multiple lesions, marginal dose was modified with consideration of total tumor volume as well as volume of the individual lesion. In repeated radiosurgery marginal dose was reduced about 25% from the dose used for the same volume in initial treatment. The mean total tumor volume and mean number of lesions at the second GKS were 7.6 cm3 (0.8–25) and 4.5 cm3 (1–22), respectively. The mean volume of the individual lesion was 2.2 cm3 (0.1–20.5) and the mean marginal dose was 15.3 Gy (12–22).
Variables
Number of patients
Age
Mean age 55.9 (range 23–78)
Sex M F
110 (54%) 94 (46%)
Primary site Lung Breast Kidney Colorectum Others
93 (46%) 26 (13%) 20 (10%) 14 (7%) 51 (25%)
WBRT Yes No
135 (66%) 69 (34%)
RPA class I II III
44 (22%) 95 (47%) 65 (32%)
Number of lesion 1–3 >3
134 (66%) 70 (34%)
Statistical analysis was performed using commercial software SPSS Version 11.0 (SPSS Inc., Chicago, IL). The Kaplan–Meier method was used to calculate the patient survival period and actuarial survival rate. The prognostic factors were analyzed for significance using univariate analysis with log-rank tests. Subsequently, a multivariate analysis was performed using Cox proportional hazard regression analysis.
Tumor volume (cm3 ) <8 >8
122 (60%) 82 (40%)
3. Results
Extracranial metastasis Yes No Primary tumor controlled Yes No
2.3. Statistical analysis
106 (52%) 98 (48%) 113 (55%) 91 (45%)
3.1. Clinical outcome in the entire group The mean follow up period was 38 weeks (4–182). A hundred fifty four patients died and 50 patients were still alive at the time of the last follow-up. The median survival of all
134
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
Fig. 1. Kaplan–Meier survival curve showing overall survival for the entire group after first radiosurgery.
the patients after initial radiosurgery was 36 (4–132) weeks (Fig. 1). The actuarial survival rate was 54% at 6 months and 32% at 1 year. The local control rate and rate without development of new lesions after the first GKS were 94.1% and 89.2% at 6 months and 86.2% and 78.3% at 1 year, respectively. The cause of death was determined in 123 patients, but was indeterminate in 31 patients. Eighty-five patients (55%) died of systemic causes, 38 (25%) of neurologic causes, and 31(25%) of unknown causes. Symptomatic radiation necrosis developed in 3 (1.5%) of the 204 patients after the first GKS. They had undergone prior WBRT with total dose of 30 Gy in 10 fractions. Tumor volume and marginal dose (cm3 /marginal dose, Gy) at the sites of radiation necrosis was 7.3/15, 4.5/20, and 3.0/25, each respectively. 3.2. Prognostic factors in the entire group Primary site, number of metastatic lesions, total tumor volume, WBRT and RPA class were included as possible prognostic factors for survival in the univariate analysis. The results are summarized in Table 2. Primary site was not a significant factor influencing survival. Number of lesions 3 or less (p = 0.001), total tumor volume 8 cm3 or less (p = 0.008), better RPA class (P < 0.001), and combined WBRT (p = 0.005) were related to longer survival. Multivariate analysis revealed that RPA class was the most dominant prognostic factors affecting survival (median survival of 64, 43, and 16 weeks in RPA classes I–III, respectively, p < 0.001) (Fig. 2).
Fig. 2. Kaplan–Meier survival curve showing overall survival stratified by recursive partitioning analysis class (A) and number of lesions (B) for all patients after radiosurgery.
3.3. Clinical outcome in salvage GKS group A subset of 43 patients (21% of 204 patients) underwent repeated GKS for local or distant recurrence (Table 3). The probability of additional procedures increased with longer survival and was 35% for those who survive longer than 6 months after the initial treatment. The mean follow-up period after the second GKS was 38 weeks (7–156). Thirty-three of these patients died and 10 were still alive at the time of the last follow-up. The mean interval between the first GKS and salvage GKS was 37 (7–80) weeks. The median survival from the first GKS was 68 (16–156) weeks in this patients group and that after the second GKS was 32 (7–132) weeks (Fig. 3). The cause of death was determined in 30 patients, but was indeterminate in three patients. Sixteen (48.4%) patients died of systemic causes and 14 (42.4%) died of neurologic causes.
Table 2 Uni- and multivariate analysis of prognostic factors for all patients Variables
Median survival (weeks)
Univariate (p-value)
Multivariate (p-value)
Primary site (lung vs. non-lung) No. of lesion (≤3 vs. >3) Tumor volume (≤8 cm3 vs. >8 cm3 ) Presence of WBRT (yes vs. no) RPA classes (I, II vs. III)
40 vs. 28 49 vs. 20 45 vs. 22 45 vs. 23 64, 43 vs. 16
0.378 0.001 0.008 0.005 <0.001
EA 0.009 0.068 0.029 <0.001
WBRT: whole brain radiotherapy; RPA: recursive partitioning analysis; EA: excluded from analysis.
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
135
Table 3 Clinical characteristics of patients in salvage radiosurgery group Variables
Number of patients
Age
Mean age 54.7 (range 35–75)
Sex M F
25 (58%) 18 (42%)
Primary site Lung Breast Kidney Others
22 (51%) 5 (12%) 7 (16%) 9 (21%)
RPA class I II III
7 (16%) 25 (58%) 11 (26%)
Number of lesion 1–3 >3
21 (49%) 22 (51%)
Tumor volume (cm3 ) <8 >8
21 (49%) 22 (51%)
Type of recurrence Local recurrence New lesion at remote site Both
16 (37%) 13 (30%) 14 (33%)
Fig. 4. Kaplan–Meier survival curve showing overall survival stratified by recursive partitioning analysis class for salvage radiosurgery group.
tumor volume change in the patients with repeated radiosurgery showed that radiosurgery for local recurrence was effective as well as initial radiosurgery in newly diagnosed patients (Fig. 5). Symptomatic radiation necrosis developed in eight patients (18.6%) after the second GKS. Mean tumor volume/marginal dose at the site of radiation necrosis was 5.8 cm3 (1.2–12.2)/19 Gy (17–20 Gy) at the first radiosurgery and 2.3 cm3 (0.2–8.4), 17 Gy (13–20) at the second radiosurgery. 3.4. Prognostic factors in salvage GKS group In the univariate analyses, interval between salvage GKS and the initial GKS, number of lesions, tumor volume, and type of recurrence (local progression or development of new lesions) were not significant factors influencing survival (Table 4). There was a significant difference in the median survival in relation to RPA class (median survival of RPA I–III was 42, 36 and 12 weeks, p = 0.007) (Fig. 4). In the multivariate analyses, the RPA class was the only statistically significant factor (p = 0.015).
Fig. 3. Kaplan–Meier survival curve showing overall survival for salvage radiosurgery group after second radiosurgery.
The actuarial survival rate after the second GKS was 57.6% at 6 months and 28% at 1 year. The local control rate and rate without development of new lesions at 6 months after the second GKS were 90.7% and 86%, respectively. The
4. Discussion Radiosurgery is recognized as a powerful treatment modality for selected patients with brain metastases [13]. It was reported that the overall median survivals ranged from 24 to 44 weeks in patients treated with radiosurgery for brain metastases of all origins [4,9,18–21]. Also, it was reported
Table 4 Uni- and multivariate analysis of prognostic factors for salvage radiosurgery group Variables
Median survival (weeks)
Univariate (p-value)
Multivariate (p-value)
Number of lesion (≤3 vs. >3) Tumor volume (≤8 cm3 vs. >8 cm3 ) Interval between the first and second GKS (≥35 weeks vs. <35 weeks) Type of recurrence (local, new lesion vs. both) RPA classes (I, II vs. III)
36 vs. 28 36 vs. 28 36 vs. 28 36, 26 vs. 32 42, 36 vs. 12
0.725 0.074 0.650 0.858 0.007
0.526 0.657 0.424 EA 0.015
RPA: recursive partitioning analysis; EA: excluded from analysis.
136
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
to be beneficial in the end-stage management of carefully selected patients with numerous brain metastases, even in those with 10 or more lesions [13–17]. However, little is known about the outcome of radiosurgery as a salvage therapy. Chen et al. [6] reported that radiosurgical salvage therapy extended the survival time. They analyzed the patients who underwent salvage radiosurgery for newly developed lesions after initial radiosurgery. In their series, the median time from the first radiosurgical procedure to the first salvage treatment was 17.4 weeks and the median survival from the second radiosurgical intervention was 28 weeks. Our study shows somewhat longer interval of 37 weeks from the first radiosurgical procedure to the first salvage treatment and a comparable median survival of 32 weeks from the second radiosurgical intervention. The median survival after the salvage GKS (32 weeks) is comparable with the overall median survival from the first GKS in all patients (36 weeks). Also the median survival after the salvage GKS for each RPA class is similar to the reported outcome of radiosurgery alone for newly diagnosed brain metastasis in the literature [22], in which the median survival for patients of RPA classes I–III were 13.4, 9.3, and 1.5 months. The median survival of the salvage treatment group from the first procedure (68 weeks) is significantly longer than that of entire group (36 weeks). One possible explanation is that the patients in the salvage treatment group have favorable prognostic factors and live long enough to experience recurrence and become candidates for repeat radiosurgery. Such a self-selection process may be an important cause of the observed longer overall survival time. However, our data also provide a reasonable interpretation for the beneficial effect of salvage GKS. Because the probability of local tumor control after salvage GKS is comparable to that after initial GKS, it is thought that multiple radiosurgery procedures result in actual prolongation of freedom from progression of both already treated lesions (shown in Fig. 5) and new lesions, thereby extending the lifetime of the patient.
Though direct comparison with other treatment is difficult due to limited data, the survival after the salvage radiosurgery in our study appears to be better than reported outcome after other method of reirradiation. In a report of whole or partial brain reirradiation [23], in which median dose of 20 Gy in 10 fractions was given for progressive brain metastases after the initial course of WBRT, overall median survival was 4 months. Though it seemed to include more patients with poor performance status at the time of reirradiation (ECOG performance status of 3 or worse in 49%) than ours at the time of the second GKS (RPA class 3 in 24%), better ECOG performance status was not associated with improved survival in that study. Overall median survival of 4 months is comparable to that of RPA class III (12 weeks) in our study and much shorter than those of RPA I (42 weeks) and II (36 weeks). In our study, the common significant prognostic factor affecting survival after the initial and salvage GKS is RPA class. It is thought that performance status becomes more influential on survival when control of brain lesion is effective. With respect to the influence of the number of lesions on survival, other authors could find no effect when treating up to four to five tumors [24–26]. The number of lesions shows a wider variation in our series, and the number of lesions or lesion volume affected survival. However, RPA class gave the most dominant influence. Therefore, it is suggested that a limited number of prognostic factors, such as the KPS score or RPA class at the time of the procedure, constitute the major determinants of outcome in both salvage GKS and initial GKS, and that the decision for additional radiosurgery should be made on the consideration of these predominant factors. Also, it should be noted that situation requiring multiple GKS procedures is not exceptional, but rather common and predictable in the course of treating metastatic brain lesions. The incidence of complications tends to increase after multiple GKS procedures, however, it seems that this risk is outweighed by the expectable prolongation of survival. Additional benefit of repeated radiosurgery may be an improvement in the quality of life, however, it is beyond the collected data of this study and further prospective trials are needed to clarify this issue.
5. Conclusion
Fig. 5. Graph depicting change of tumor volume in patients with locally recurrent lesions after the first and second GKS. Data are from 16 patients in whom MR images were obtained at 3 months after the second GKS. The volume of recurrent lesions was stable or decreased in 13 patients and increased in three after repeated treatment. Tumor volume is presented with mean percentage of the volume at the first GKS and standard deviation. * Percentage volume at the time of maximal reduction during serial followup.
Local or remote recurrence which needs salvage treatment is common for patients with metastatic brain tumors after initial radiosurgery. Repeated GKS is effective in treating locally progressive or new lesions after initial GKS. If the major prognostic factors such as RPA class are favorable, repeated radiosurgical procedure may be beneficial for an increase in survival as like initial radiosurgery.
Acknowledgment This work was supported by a grant of Insung Foundation for Medical Research (J.-I. Lee).
K.-Y. Kwon et al. / Clinical Neurology and Neurosurgery 109 (2007) 132–137
References [1] Adler JR, Cox RS, Kaplan I, Martin DP. Stereotactic radiosurgical treatment of brain metastases. J Neurosurg 1992;76:444–9. [2] Alexander III E, Moriarty TM, Davis RB, Wen PY, Fine HA, Black PM, et al. Stereotactic radiosurgery for the definitive, noninvasive treatment of brain metastases. J Natl Cancer Inst 1995;87:34–40. [3] Auchter RM, Lamond JP, Alexander E, Buatti JM, Chappell R, Friedman WA, et al. A multiinstitutional outcome and prognostic factor analysis of radiosurgery for resectable single brain metastasis. Int J Radiat Oncol Biol Phys 1996;35:27–35. [4] Flickinger JC, Kondziolka D, Lunsford LD, Coffey RJ, Goodman ML, Shaw EG, et al. A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys 1994;28:797–802. [5] Shirato H, Takamura A, Tomita M, Suzuki K, Nishioka T, Isu T, et al. Stereotactic irradiation without whole-brain irradiation for single brain metastasis. Int J Radiat Oncol Biol Phys 1997;37:385–91. [6] Chen JCT, Petrovich Z, Giannotta SL, Yu C, Apuzzo ML. Radiosurgery salvage therapy for patients presenting with recurrence of metastatic disease to the brain. Neurosurgery 2000;46:860–7. [7] Chen JCT, Petrovich Z, O’Day S, Morton D, Essner R, Giannotta SL, et al. Stereotactic radiosurgery in the treatment of metastatic disease to brain. Neurosurgery 2000;47:268–81. [8] Gerosa M, Nicolato A, Foroni R, Zanotti B, Tomazzoli L, Miscusi M, et al. Gamma knife radiosurgery for brain metastases: a primary therapeutic option. J Neurosurg 2000;97(Suppl 5):515–24. [9] Kim YS, Kondziolka D, Flickinger JC, Lunsford LD. Stereotactic radiosurgery for patients with nonsmall cell lung carcinoma metastatic to the brain. Cancer 1997;80:2075–83. [10] Gaspar L, Scott C, Rotman M, Asbell S, Phillips T. Recursive partitioning analysis of prognostic factors in three radiation therapy oncology group brain metastases trials. Int J Radiat Oncol Biol Phys 1997;37(4):745–51. [11] Therasse P, Arbuck SG, Eisenhauer EA, Wanders J. New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 2000;92(3):205–16. [12] Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90-05. Int J Radiat Oncol Biol Phys 2000;47(2):291–8. [13] Yamamoto M, Ide M, Nishio. Gamma knife radiosurgery for numerous brain metastases: is this a safe treatment? Int J Radiat Oncol Biol Phys 2002;53(5):1279–83.
137
[14] Suzuki S, Omagari J, Nishio S, Nishiye E, Fukui M. Gamma knife radiosurgery for simultaneous multiple metastatic tumors. J Neurosurg 2000;93(Suppl 3):30–1. [15] Nam TK, Lee JI, Jung YJ, Im YS, An HY, Nam DH, et al. Gamma knife surgery for brain metastases in patients harboring four or more lesions: survival and prognostic factors. J Neurosurg 2005;(Suppl 102):147– 50. [16] Yamamoto H, Takagi K, Ushio T. Gamma knife radiosurgery for diffuse brain metastases. Noushinkeigeka Sokuho 1998;8:391–8. [17] Yamamoto M, Ide M, Jimbo M. Gamma knife radiosurgery with numerous target points for intracranially disseminated metastases: early experience in 3 patients and experimental analysis of whole brain irradiation doses. Radiosurgery 1997:94–109. [18] Engenhart R, Kimmig BN, Hover KH, Hover KH, Wowra B, Romahn J, et al. Long-term follow-up for brain metastases treated by percutaneous stereotactic single high-dose irradiation. Cancer 1993;71:1353– 61. [19] Kihlstrom L, Karlsson B, Lindquist C. Gamma knife surgery for cerebral metastases. Implications for survival based on 16 year experience. Stereotact Funct Neurosurg 1993;61(Suppl 1):45–50. [20] Kim DG, Chung HT, Gwak HS, Paek SH, Jung HW, Han DH. Gamma knife radiosurgery for brain metastases: prognostic factors for survival and local control. J Neurosurg 2000;93(Suppl 3):23– 9. [21] Somaza S, Kondziolka D, Lunsford LD, Kirkwood JM, Flickinger JC. Stereotactic radiosurgery for cerebral metastatic melanoma. J Neurosurg 1993;79:661–6. [22] Lutterbach J, Cyron D, kenne K, Ostertag CB. Radiosurgery followed by planned observation in patients with one to three brain metastases. Neurosurgery 2003;52:1066–74. [23] Wong WW, Schild SE, Sawyer TE. Analysis of outcome in patients reirradiated for brain metastases. Int J Radiat Oncol Biol Phys 1996;34(3):585–90. [24] Chen J, O’Day S, Morton D, Essner R, Cohen-Gadol A, Macpherson D, et al. Stereotactic radiosurgery in the treatment of metastatic disease to the brain. Neurosurgery 2000;47(2):268–79. [25] Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC. Stereotactic radiosurgery plus whole radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat Oncol Biol Phys 1999;45:427–34. [26] Sneed P, Lamborn K, Forstner J, McDermott M, Chang S, Park E, et al. Radiosurgery for brain metastases: Is whole brain radiotherapy necessary? Int J Radiat Oncol Biol Phys 1999;43:549–58.